Brain Mechanisms

Dividing up aggression and considerations in studying the physiological substrates of these phenomena. I will forego commentary on the topography of the neural circuits that underlie aggressive phenomena - these will, no doubt, occupy other commentators. I have devoted considerable effort (Brain 1977; 1978; 1979) to creating a meaningful classification for the diverse phenomena included in the concept of "aggression" as it is applied to infrahumans. Again, the intention was ultimately to relate behavior to physiology. Adams's account is particularly useful in this respect. His categories seem likely to have wide application. For example, offensive and defensive forms of attack have been distinguished (on the basis of bite targets) in laboratory mice (Childs and Brain 1979a; b).

I propose to ask three basic groups of questions of this scheme. The first group concerns whether division of all intraspecific conflict behaviors into offense, defense, and subordination provides one with sufficient information for the eventual identification of the physiological bases ot these activities? Adams's basic separation is logical and defensible, but, are the categories too wide, and do they all refer to aggression? One may note that, while submission is clearly an important component of agonistic behavior (used in the ethological sense), it is not aggression (in the sense of relating to phenomena that involve the "intentional" direction of "noxious stimuli" towards conspecifics). It would not diminish the importance of submission to maintain that it is part of a continuum that incorporates offense. The distinction between offense and defense seems superficially clear, but aggressive encounters (which generally involve more than one animal) may involve fluctuations between activities that can be labeled as "offense, defense, and submission." Are defensive postures preliminaries to defensive attack or are they ambivalent components of offense? I concur with Adams that one should attempt to identify the major motivation involved in aggression, but I am less optimistic about their clear separation. Are separate neural mechanisms necessary, or even likely?

I would have liked Adams to have considered the neural basis of predatory responses in cats and rats. Although he dismissed them as being (largely) interspecific reactions, they do feature prominently in research on the neural control of aggression. The following categories of aggression may provide more information than Adams's scheme:

1. Self-defense ("fear"-mediated attack related to pain and escape).
2. Maternal aggression (attack in situations where litters, nest sites, etc, are threatened).
3. Predatory aggression (if it can be called aggression!).
4. Reproduction termination (in mammals this is the pup-killing response).
5. Social aggression (offensive attack, generally related to competition for mates, territories, or social status).

The first two are defense in Adams's scheme and the last is offense. The dubious nature of categories 3 and 4 has been emphasized elsewhere (Brain 1979), but they are currently used as models of "aggression."

A second group of questions concerns hormonal involvement in the phenomena described by Adams. Adams did not emphasize that endocrine influences on attack/defense/subordination depend on the species, the behavioral context, the particular hormone, its dose, temporal relationships, and prior experience. For example, the statement that testosterone mediates offensive attack in these species is a gross simplification. Neural conversion of androgens may be implicated in some species (Bowden and Brain 1976; Brain and Bowden 1978). In other situations, offense seems inversely related to testosterone titer: group- housed, castrated mice show increased attack (compared to intact counterparts) on intruding lactating females (Haug and Brain 1976). This behavior is suppressed by testosterone application. Pheromones (Brain and Haug, in press) mediating this response cannot be "dependent on testosterone production." Adams's models may benefit by considering (a) neural locations at which steroid-hormones can be shown to be autoradiographically concentrated, and (b) sites at which intracranially-administered hormones prove particularly effective, For example, the cited glucocorticoid influences on defense may be mediated by the postulated hippocampal modulation of septal activity, as the hippocampus is a prime target for these hormones.

The final fundamental questions relate to Adams's statements on the evolution of aggression. The view that defensive responses are derived from antipredatory reactions reminds one of Huntingford's (1976) distinction among "social aggression," "predation," and "antipredatory" reactions, Social aggression (perhaps the defensive components of these activities?) was said to have greater affinity with antipredatory responses than with predatory ones. This view is still highly controversial however (see Brain 1979), Some statements in the Adams account are circuitous, such as "the submission system evolved later to modify defense behavior when the animal was confronted with a conspecific whose offense behavior could be inhibited by particular submissive postures." The hope that "the primates can tell us something of the evolution of brain mechanisms for aggression in humans" is likely to prove particularly provocative. Do the similarities between the neuroanatomical areas that are involved in conflict in the small number of investigated species, imply that mammals have inherited "aggression circuits?" The alternative view is that aggression (in its various forms) is best viewed as a group of strategies that may be independently acquired by different species (see Eibl-Eibesfeldt 1977). Superficial similarities in neuroendocrine control may be generated by common requirements and limited possibilities; for example, as males are often required to show enhanced aggression in competition for breeding resources, a modulation by "androgens" and particular neural areas becomes especially likely. More obviously, as mammals have limited weapons of offense and defense (teeth, claws, etc,), this is likely to be reflected in similar motor control of aggression. More species and situations need to be systematically evaluated before one can chose between these alternatives.

I applaud the tone and content of Adams's paper. It provides a fair evaluation of some of the problems inherent in studies on the physiological control of aggression. There will no doubt be fierce debate concerning the existence and form of the neuro-endocrine mechanisms that motivate these activities, but it would help enormously if workers would (like Adams) pay more attention to behavioral intricacies. Slack use of the term aggression has generated enormous confusion.